The present application relates to a method for the prognosis of adverse outcomes in a subject comprising the following steps: (i) determining the midkine level in a sample of said subject； and (ii) comparing the determined midkine level to a control midkine level； wherein said sample is taken after the subject has been subjected to cardiovascular stress； wherein said control midkine level is derived from control sample(s) from one or more subjects not showing adverse outcomes, wherein the control sample(s) have been taken after said one or more subjects not showing adverse outcomes have been subjected to cardiovascular stress； and wherein a decreased determine midkine level as compared to the control midkine level is indicative of an adverse outcome. Further it relates to a method for the prognosis of adverse outcomes in a subject comprising the following steps: (i) determining in samples of said subject the increase in midkine levels (Δmidkine value) during cardiovascular stress； and (ii) comparing the determined Δmidkine value to a control Δmidkine value； wherein said control Δmidkine value is derived from one or more subjects not showing adverse outcomes； wherein a decreased determined Δmidkine value as compared to the control Δmidkine value is indicative of an adverse outcome. It also relates to the use of a midkine antibody or an antigen binding fragment thereof for the prognosis of adverse outcomes in a subject undergoing dialysis therapy. Also encompassed is midkine for use in the treatment or the prevention of an adverse outcome, wherein midkine is administered to the subject.La présente invention concerne un procédé de pronostic d'évolutions défavorables chez un sujet comprenant les étapes suivantes consistant : i) à déterminer le niveau de midkine dans un échantillon dudit sujet ； et (ii) à comparer le niveau déterminé de midkine à un niveau de midkine témoin ； ledit échantillon étant prélevé après que le sujet a été soumis à un stress cardio-vasculaire ；
